Buckyplates and Buckybowls: Examining the Effects of Curvature on Ï

Article pubs.acs.org/JPCA

Buckyplates and Buckybowls: Examining the Effects of Curvature on π−π Interactions Matthew R. Kennedy, Lori A. Burns, and C. David Sherrill* Center for Computational Molecular Science and Technology, School of Chemistry and Biochemistry, and School of Computational Science and Engineering, Georgia Institute of Technology, Atlanta, Georgia 30332-0400, United States S Supporting Information *

ABSTRACT: π−π interactions are integral to many areas of chemistry, biochemistry, and materials science. Here we use electronic structure theory to analyze how π−π interactions change as the π-systems are curved in model complexes based on coronene and corannulene dimers. Curvature redistributes electronic charge in the π-cloud and creates a dipole moment in these systems, leading to enhanced intermolecular electrostatic interactions in the concave-convex (nested) geometries that are the focus of this work. Curvature of both monomers also has a geometric effect on the interaction by decreasing the average C−C distance between monomers and by increasing the magnitude of both favorable London dispersion interactions and unfavorable exchange-repulsion interactions. Overall, increasing curvature in nested π−π interactions leads to more favorable interaction energies regardless of the native state of the monomers, except at short distances where the most highly curved systems are less favorable as exchange repulsion terms begin to dominate the interaction.

■

INTRODUCTION π−π interactions play a pivotal role in the way proteins fold1,2 and drugs bind to targets.3 They are also critical in determining the crystal packing and charge transport properties of πconjugated organic materials for organic electronics applications.4 In biological contexts and frequently in organic materials, π−π interactions involve planar π-systems, and the fundamental physics of planar π−π interactions has been the focus of significant recent effort.1,3,5−7 π−π interactions involving prototypical molecules such as benzene,6,8−12 coronene,13,14 and linear acenes15 have been given special consideration. However, curved π systems including fullerenes and carbon nanotubes are also of significant interest in materials applications. Experimental and theoretical works have examined such curved π systems,16−20 yet to our knowledge, the effect of curvature on the fundamental physics of π−π interactions has not been systematically explored. To investigate this issue, we employed as model systems nested dimers of coronene and corannulene, which are among the simplest model systems large enough to exhibit curvature effects. Coronene dimer is naturally flat because its central ring has six sides, and corannulene is naturally curved because its central ring has five sides. High-quality interaction energies for coronene dimer,13 corannulene dimer,21 and coronene− circumcoronene dimer14 have recently been published by Pulay and co-workers. To probe the effect of curvature directly, we constructed model dimer geometries in which each monomer (coronene and corannulene) was distorted from its © 2012 American Chemical Society

natural state (flat and curved, respectively) through to its opposite state (curved and flat). Using these model systems, we are able to examine the interaction between monomers (at fixed monomer geometries) as a function of intermonomer distance and monomer curvature. Curvature can affect the intermolecular interaction through two routes: by modifying the electronic structure of the monomers and by introducing differences in the contact geometries. An example of the former is that as the monomers become curved, a permanent dipole is created within each monomer that interacts with the dipole on the other monomer for an enhanced or reduced electrostatic interaction, depending on the orientation of the two monomers. Here we seek to quantify such effects to better understand how curvature influences the strength and character of π−π interactions. Dispersion plays a major role in π−π interactions, and hence we also use these systems to examine possible differences between recently proposed dispersion corrections to density functional theory. The simplest such corrections are damped C6R−6 terms,22 popularized primarily by Grimme.23−25 In most such approaches (including Grimme’s second-generation -D2 method24), the C6 parameters are tabulated for each element and are not able to respond to their local chemical environment (and thus, for present purposes, not able to adapt to curving Received: June 11, 2012 Revised: October 30, 2012 Published: November 8, 2012 11920

dx.doi.org/10.1021/jp305700k | J. Phys. Chem. A 2012, 116, 11920−11926

The Journal of Physical Chemistry A

Article

Figure 1. (a) Side view of coronene dimer geometry with τ = 50° and R = 3.7 Å. (b) Side view of coronene dimer geometry with τ = 0° and R = 3.7 Å. (c) Top-down view of coronene dimer geometry where a, b, and c label distinct groups of carbon atoms. The curvature τ and the dihedral τ′ are related by τ′ = 180 − (τ/2).

monomers). However, of course the C6 parameters should change with curvature, and here we examine by how much, using two alternative approaches. First, we employ Grimme’s third-generation-D3 method,25 which interpolates C6 terms from tabulated data based on each atom’s coordination number. Second, we employ a more theoretically sophisticated approach, Becke and Johnson’s exchange-hole dipole moment model (XDM).26−28 Johnson has recently shown that oxidation state, hybridization, charge, and other chemical properties can affect the atomic Cn by as much as 65% for carbon atoms.29

relaxation energies are not particularly relevant to our studyour focus is on how intermolecular interaction energies (at fixed monomer geometries) are affected by curvature, regardless of how the monomer curvature might have been achieved (the curvature will be natural in our corannulene dimer, or in larger systems like carbon nanotubes or fullerenes). Dispersion-Corrected DFT. Single-point energy computations were performed using the B3LYP,33 B97-D,24 M05-2X,34 and M06−2X35 functionals with the heavy-aug-cc-pVDZ (haDZ) basis set (defined as aug-cc-pVDZ36 for non-hydrogen atoms and cc-pVDZ37 for hydrogen atoms) to determine which functional most accurately reproduced the estimated triple-ζ quality QCISD(T) interaction energies for concave−convex stacked corannulene dimer recently reported by Janowski et al.21 That work examined the B97-D, M06−2X, and ωB97X-D functionals in conjunction with the cc-pVQZ basis set and noted that B97-D provided an intermonomer distance within 0.06 Å and an interaction energy within 1.1 kcal mol−1 of benchmark values. DFT computations reported in this work utilized NWChem 6.0,38 with the “fine” or (70,590) grid. Grimme’s -D2 and -D3 corrections24,25 were both applied to all of the functionals through the dftd3 program.25 The computations of XDM dispersion coefficients were performed in Q-Chem 3.239 with the full aug-cc-pVDZ basis and damped into a dispersion correction according to revised parameters.40 Generally our computations are counterpoise corrected according to the Boys and Bernardi approach41 to reduce basis set superposition error (BSSE); however, the XDM computations are not counterpoise corrected because parameters for the XDM damping function were optimized using uncorrected interaction energies. The three dispersion corrections considered here all take slightly different forms. Grimme’s DFT-D2 correction24 is the simplest, employing C6 only in the dispersion series, and is expressed in eq 1.

■

COMPUTATIONAL METHODS Geometries. To investigate the effect of curvature on π−π interactions, a set of coronene (buckyplate) and corannulene (buckybowl) dimers was constructed by varying the intermonomer distance and the degree of curvature of the monomers. The labeling of atoms, bonds, and curvature angles for coronene can be seen in Figure 1 and is analogous for corannulene. The curved monomer geometries were constructed such that bond lengths were fixed while bond angles changed in order to relieve the stress of curvature. The formulae for construction of the molecules can be found in the Supporting Information. In order to find a reliable, computationally tractable methodology for treating these systems, computations using four density functionals and four different dispersion corrections were performed on the geometries of corannulene dimer reported by Janowski et al.21 These structures are nearly identical to those obtained by X-ray crystallography,30 with curvature torsion angle (see Figure 1) of τ ≈ 57.5°, and were obtained using the PBE1 functional31 with the 6-31G* basis set32 using a grid with 99 radial shells and 302 angular points (99,302). Here we employ frozen monomer geometries to simplify the computations. This approximation appears to be a good one for these systems,21 at least for native monomer curvatures. Of course, there would be large monomer relaxation energies for our artificially curved monomers, but given the artificial nature of the imposed curvature, the associated 11921


The Journal of Physical Chemistry A N D2 Edisp = −s6

∑ i,j>i

Article

and the fluctuation potential (electron correlation), as explained in ref 49. Examining the components along a potential energy curve provides insight into how the nature of the interaction changes as the monomers approach each other. These calculations were performed with the jun-cc-pVDZ (jaDZ) basis set51 (formerly referred to by our group as aug-ccpVDZ′) inside a development version of PSI4.52 Compared to the aug-cc-pVDZ basis, jun-cc-pVDZ removes diffuse functions from hydrogen atoms and diffuse d functions from the nonhydrogen elements. For SAPT0, jun-cc-pVDZ has been shown to yield favorable error cancellation.53

C6ij

f (R ij) 6 damp

R ij

(1)

The C6 coefficients are calculated as the geometric average between the two elemental C6 coefficients Ci6 and Cj6. The s6 parameter is optimized for each functional and is usually between 0.5 and 1.5. Grimme recently released a more refined method,25 DFTD3, which additionally implements the R−8 term in the dispersion correction, changes the method of calculating C6, and alters the damping function. Atomic Ci6 coefficients are interpolated from reference values based upon coordination number in the current molecular structure. Thus, the C6 values are allowed to respond to the molecular environment in -D3, rather than simply being looked-up from a table of elemental values. The DFT-D3 correction has the form: N D3 Edisp =−

∑ n = 6,8

sn

∑ i,j>i

Cnij f (R ij) R ijn damp

■

RESULTS AND DISCUSSION The performance of density functional methods versus the estimated QCISD(T)/aug-cc-pVTZ benchmarks for corannulene dimer reported by Janowski et al.21 is presented in Figure 2

(2)

We investigated two different damping functions for use with the DFT-D3 correction. The first is the original damping function for DFT-D3, which damps the dispersion interaction to zero at short ranges,25 and the other is the damping function suggested by Becke and Johnson,26,42,43 and now advocated by Grimme, DFT-D3(BJ). Implementing the latter entailed a reoptimization of the sn parameters as well as a switch of damping functional form.44 The last dispersion correction considered is the XDM method of Becke and Johnson.26,28,45,46 Calculation of the XDM correction is based upon the nonzero dipole moment of exchange holes47 on one monomer that can induce a complementary dipole moment on the other monomer, leading to a dispersion interaction between the molecules. This method is quite different from the first two approaches to dispersion corrections because XDM calculates Cn parameters from the electron density instead of using tabulated C6 data. The computation of the XDM dispersion contribution to the energy in eq 3 uses a similar functional form as those used for the DFT-D2 and DFT-D3 corrections but involves one term further, C10. N XDM Edisp =−

∑ ∑ n = 6,8,10 i , j > i

Cnij f (R ij) R ijn damp

Figure 2. Counterpoise-corrected DFT/haDZ interaction energies for corannulene dimer compared to QCISD(T) reference values from ref 21. (See text for XDM details).

as a function of intermonomer distance. The benchmark curve shows a minimum of 15.43 kcal mol−1 at 3.64 Å. It is apparent that M05-2X and M06−2X both underbind (M05-2X by a larger margin) when compared to the reference, with neither curve tracking the shape of the reference curve very closely. The B3LYP-D3 and B97-D3 methods also underbind the reference by 3−4 kcal mol−1. B3LYP-D2 and B97-D2 are closer to the reference data than their respective -D3 curves, but both have the wrong shape, shifting the equilibrium distance to shorter distances. B97-D3(BJ) and B3LYP-D3(BJ) both show good agreement with the shape of the reference curve, and the B3LYP-D3(BJ) curve is rather close to the benchmark curve (differences of 0.4−0.8 kcal mol−1). The last method investigated is B3LYP-XDM using the parametrization of Burns et al.,40 which performs very well in providing a curve with the right shape; the interaction energies are underestimated by 2.0−2.5 kcal mol−1 across the potential curve. However, these errors are significantly larger than the errors exhibited by B3LYP-D3(BJ). Hence, we selected the latter method for primary use in this work. We also performed limited tests using XDM with the PBE functional, but the dimers were more underbound using this functional than they were using B3LYP-XDM. Results for the full two-dimensional scan of intermonomer distance and curvature (R, τ) are displayed in Figure 3. The B3LYP-D3(BJ)/haDZ computations show that as τ is varied from 0° to 50°, minima of the coronene interaction energy curves range from −17.7 to −22.9 kcal mol−1 (−13.1 to −16.2

(3)

To help illustrate electronic changes due to curvature, electrostatic potentials of coronene and corannulene at different curvatures were produced using the Spartan ’08 package;48 the electrostatic potential [between −80 (red) and +80 (blue) kcal mol−1] was mapped onto the electron density (produced at the B3LYP/6-31+G level of theory) with an isovalue of 0.002 e− bohr−3. Our discussions of the plotted electrostatic potentials are purely qualitative, and we do not expect the figures to change noticeably at higher levels of theory.

■

SAPT ENERGY DECOMPOSITION Symmetry adapted perturbation theory (SAPT) is an approach by which individual physical components (electrostatics, induction, dispersion, and exchange) of a nonbonded interaction can be elucidated.49,50 SAPT partitions the Hamiltonian into contributions from the Fock operator of each monomer, the interaction between the two monomers, 11922



Article

2.5 D at τ = 50°. Counterbalancing these increases in attractive components with increasing curvature is an increase in the unfavorable exchange-repulsion term. Of the two primary contributors to extra stabilization with increasing curvature (electrostatics and London dispersion forces), let us begin with the electrostatic term. The electrostatic interaction becomes more favorable with curvature in these concave−convex (nested) geometries for two reasons: first, the curvature redistributes the electronic charge leading to more favorable interactions, and second, at a fixed distance between the central rings, greater curvature leads to shorter average intermolecular C−C distances, in turn leading to larger favorable charge penetration terms. The first effect, that coming from charge redistribution, is illustrated in Figures 5 and 6. For flat monomers, the two sides of the monomer are symmetric, and hence the charge distribution is equivalent on both sides. However, upon curvature, the monomers adopt a positive charge on the outside of the bowl and a negative charge on the inside, leading to an electronic dipole moment. For the nested (concave− convex) geometries considered here, this means that the two dipoles are aligned in a head-to-tail arrangement, which contributes to a favorable electrostatic interaction. The dipole moment of the monomers is a significant 2.5 D when τ = 50°. The second effect, a geometric effect in which average interatomic contact distances get shorter, also enhances the electrostatic interaction. For a fixed distance between the central rings of the two monomers, curvature of both monomers means that the atoms around the periphery of the outer monomer get closer to the atoms of the inner monomer in a concave−convex geometry. As the curvature changes from 0° to 50° at an intermolecular distance of 3.7 Å, this results in a 10% reduction in the sum of interatomic distances between pairs of atoms on separate monomers. Having closer contact between the monomers is favorable electrostatically because it leads to enhanced charge penetration terms. Charge penetration is a favorable contribution to electrostatic interactions due to orbital overlap, and its importance to π−π interactions has recently been emphasized by Hohenstein et al.56 The effect is visible in the SAPT0 electrostatic term plotted in Figure 4, which becomes both exponentially more attractive at short intermonomer distances and more attractive with increasing curvature for a fixed intermonomer distance. A comparison of the more rigorous (charge-penetration including) SAPT electrostatic energy and a (charge-penetration neglecting) distributed multipole analysis57 (DMA) estimate is presented in Figure 7. The DMA is performed based on

Figure 3. B3LYP-D3(BJ)/haDZ interaction energy curves for various curvatures of coronene dimer (top panel) and corannulene dimer (bottom panel).

kcal mol−1 for corannulene). Coronene and corannulene exhibit the same trend from 0° to 50°: the minima of the interaction energy curves deepen with increasing curvature. This stabilization results from an enhancement in both the dispersion interaction and electrostatic interaction between the two monomers, as seen from the SAPT0 results plotted in Figure 4. Induction also increases with increasing curvature (Figure 4), although this effect is smaller than that for the electrostatic and dispersion terms. Induction accounts for approximately 25% of the interaction energy at the most extreme curvature studied, and 6.5% at τ = 0°. Such percentages are higher than those typically observed in other π systems that have been examined to date.54,55 The relatively large induction interaction is due to a very strong dependence of the monomer dipole moment on curvature, increasing from 0 at τ = 0° to

Figure 4. SAPT0 breakdown of the interaction energy for various values of τ. Top panel: coronene, bottom panel: corannulene. 11923



Article

Figure 5. Electrostatic potential at the B3LYP/6-31+G* level of theory, using a scale of −80 (red) to +80 (blue) kcal mol−1 for (a) coronene with τ = 0°, (b) convex side of coronene with τ = 50°, and (c) concave side of coronene with τ = 50°.

Figure 6. Electrostatic potential at the B3LYP/6-31+G* level of theory, using a scale of −80 (red) to +80 (blue) kcal mol−1 for (a) corannulene with τ = 0°, (b) convex side of corannulene with τ = 50°, and (c) concave side of corannulene with τ = 50°.

5 Å separation (where a multipole description of the monomers becomes more appropriate), the deviation is less than 1 kcal mol−1. Again, the SAPT electrostatics term shows more favorable interactions for more curved monomers at a fixed intermolecular distance. Next, let us consider changes to the dispersion contribution as a function of curvature. Figure 4 indicates that the dispersion term becomes more favorable with increasing curvature at a fixed distance between the central rings of the monomers. Just as for the electrostatic term, there can be two origins of this change: an electronic effect (changes in electron distribution and polarizability) and a geometric effect (closer interatomic contacts with curvature). Because the dispersion interaction depends sensitively on interatomic distances (the leading term being proportional to R−6), the geometric effect on dispersion is easy to understand. The -D2 dispersion contribution to the interaction energy goes from −25.69 to −33.96 kcal mol−1 for coronene as the curvature is increased from 0° to 50°, as seen in Table 1. The increase in the magnitude of the dispersion term as computed by -D2 is from geometric effects alone, because the C6 coefficients in that approximation are fixed for each element and are not dependent on chemical environment. Hence, the geometric contribution to the dispersion term is substantial. Changes in the electronic distribution with respect to curvature are another possible source of changes to the dispersion term. Grimme’s -D3 model allows for some response of the dispersion coefficients to an atom’s coordination geometry, and the XDM model allows for a full response to the chemical environment because the dispersion coefficients are computed directly from the electron density. To quantify how much the changing chemical environment affects the dispersion contribution, we evaluated -XDM and -D3(BJ) dispersion terms at 0° and 50° curvature, for both coronene dimer and corannulene dimer. Next, we took the computed dispersion coefficients Cn at each geometry, and then applied them at the other geometry to re-evaluate the dispersion

Figure 7. Comparison of SAPT0 electrostatic contribution (dashed lines) and distributed multipole analysis (DMA) of electrostatics through quadrupole−quadrupole (solid lines) at various τ. Top panel: coronene, bottom panel: corannulene.

Hartree−Fock/6-311G** results and includes terms through quadrupole−quadrupole. The DMA analysis actually shows repulsive electrostatics at short distances, indicating that electrostatics in these systems would not be favorable if not for the charge penetration terms. For dimers 3.3 Å apart, the deviation is upward of 40 kcal mol−1, whereas for dimers over 11924



Article

Table 1. Dispersion Corrections Relative to 0° Geometry with 0° Cn Coefficients at R = 3.7 Åa coronene geometry from 0° 0° 50° 50°

C6 from 0° 50° 0° 50°

-D2 0.00 0.00 −8.27 −8.27

corannulene

-D3(BJ) b

0.00 −0.61 −7.23 −6.75

-XDM

c

0.00 −0.08 −5.64 −5.34

d

-D2 0.00 0.00 −6.29 −6.29

-D3(BJ) e

0.00 −0.51 −5.21 −5.76

f

-XDM 0.00g −0.08 −4.31 −4.22

a All values in kcal mol−1. b-D2 dispersion interaction for coronene is −25.69 kcal mol−1. c-D3(BJ) dispersion interaction for coronene is −28.85 kcal mol−1. d-XDM dispersion interaction for coronene is −20.80 kcal mol−1. e-D2 dispersion interaction for corannulene is −20.63 kcal mol−1. f-D3(BJ) dispersion interaction for corannulene is −22.28 kcal mol−1. g-XDM dispersion interaction for corannulene is −16.89 kcal mol−1.

contribution. This provides a measure of how much the changing Cn coefficients change the total dispersion contribution in these systems as they progress from flat to curved. Table 1 shows dispersion interactions for the two dimers, relative to their dispersion interactions at τ = 0°. Comparing the first two rows, we see that the Cn coefficients affect the dispersion interaction of corannulene by 0.51 and 0.08 kcal mol−1 for -D3(BJ) and -XDM, respectively, when applying the Cn coefficients from τ = 50° to the geometry τ = 0°. These are modest contributions compared to the overall dispersion contributions of −16 to −29 kcal mol−1 (see Table footnote). Comparing the first and third rows of the table shows that using fixed Cn values on two different geometries changes the dispersion interaction by 5.2 or 4.3 kcal mol−1 for -D3(BJ) and -XDM, respectively, for corannulene. This is a much larger change than that observed in the previous comparison. This indicates that the geometric effect is around 10−50 times larger than the electronic effect in determining the changes in dispersion contributions with respect to curvature. Results for coronene in Table 1 show similar trends. Lastly, in Figure 8 we examine the changes of the C6 coefficients in both -D3(BJ) and -XDM for various τ. First, the coefficients from -D3(BJ) and -XDM follow the same trends with respect to monomer curvature. Second, both sets of C6 coefficients change by no more than 5% between τ = 0° and τ = 50°, consistent with the results from Table 1 discussed above.

■

CONCLUSIONS We have studied concave−convex (nested) dimers of corannulene and coronene at various imposed monomer curvatures to elucidate how curvature affects the fundamental physics of π−π interactions. Both coronene and corannulene dimers exhibit more attractive interaction energies upon curvature, except at short, unnatural intermolecular distances and high degrees of curvature. Curvature in a π system affects all four major components of noncovalent interactions: dispersion, electrostatics, exchange, and induction. Dispersion and exchange are affected most strongly by simple geometric effects because average interatomic distances between monomers decrease with increasing curvature in concave-convex geometries as the peripheral atoms of the outer monomer become closer to the atoms of the inner monomer. This same geometric effect also leads to increased orbital overlap, and hence larger stabilizing charge penetration contributions to the electrostatic interaction. However, curvature creates a charge redistribution within the monomers that also affects the nature of the interaction. XDM and DFT-D computations suggest that this electronic effect is much less important than the geometric effect when examining the dispersion term. This charge redistribution has a significant electrostatic effect because it

Figure 8. C6 coefficients from -D3(BJ) (solid lines, left axis) and -XDM (dashed lines, right axis) at various τ for the three distinct carbon atoms defined in Figure 1. Top panel: coronene, bottom panel: corannulene.

creates a dipole on each curved monomer, which leads to a favorable head-to-tail alignment of dipoles in concave-convex geometries. We expect that the careful elucidation of these various factors determining how π−π interactions vary with curvature will aid future studies that explore diverse chemical systems featuring such interactions.

■

ASSOCIATED CONTENT

S Supporting Information *

Geometrical information for the systems evaluated in this work as well as data tables for Figures 2 and 3. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. 11925



Article

Notes

(37) Dunning, T. H. J. Chem. Phys. 1989, 90, 1007−1023. (38) Valiev, M.; Bylaska, E. J.; Govind, N.; Kowalski, K.; Straatsma, T. P.; Van Dam, H. J. J.; Wang, D.; Nieplocha, J.; Apra, E.; Windus, T. L.; et al. Comput. Phys. Commun. 2010, 181, 1477−1489. (39) Shao, Y.; Molnar, L. F.; Jung, Y.; Kussmann, J.; Ochsenfeld, C.; Brown, S. T.; Gilbert, A. T. B.; Slipchenko, L. V.; Levchenko, S. V.; O’Neill, D. P.; R., A. D., Jr.; Lochan, R. C.; Wang, T.; Beran, G. J. O.; Besley, N. A.; et al. Phys. Chem. Chem. Phys. 2006, 8, 3172−3191. (40) Burns, L. A.; Vázquez-Mayagoitia, Á .; Sumpter, B. G.; Sherrill, C. D. J. Chem. Phys. 2011, 134, 084107. (41) Boys, S. F.; Bernardi, F. Mol. Phys. 1970, 19, 553−566. (42) Johnson, E. R.; Becke, A. D. J. Chem. Phys. 2005, 123, 024101. (43) Johnson, E. R.; Becke, A. D. J. Chem. Phys. 2006, 124, 174104. (44) Grimme, S.; Ehrlich, S.; Goerigk, L. J. Comput. Chem. 2011, 32, 1456−1465. (45) Angyan, J. G. J. Chem. Phys. 2007, 127, 024108. (46) Hesselmann, A. J. Chem. Phys. 2009, 130, 084104. (47) Becke, A. D.; Roussel, M. R. Phys. Rev. A 1989, 39, 3761. (48) Spartan ’08; Wavefunction, Inc.: Irvine, CA. (49) Jeziorski, B.; Moszynski, R.; Szalewicz, K. Chem. Rev. 1994, 94, 1887−1930. (50) Szalewicz, K. WIREs Comput. Mol. Sci. 2012, 2, 254−272. (51) Papajak, E.; Zheng, J.; Xu, X.; Leverentz, H. R.; Truhlar, D. G. J. Chem. Theory Comput. 2011, 7, 3027−3034. (52) Turney, J. M.; Simmonett, A. C.; Parrish, R. M.; Hohenstein, E. G.; Evangelista, F. A.; Fermann, J. T.; Mintz, B. J.; Burns, L. A.; Wilke, J. J.; Abrams, M. L.; et al. WIREs Comput. Mol. Sci. 2012, 2, 556−565. (53) Hohenstein, E. G.; Sherrill, C. D. J. Chem. Phys. 2010, 133, 014101. (54) Hohenstein, E. G.; Parrish, R. M.; Sherrill, C. D.; Turney, J. M.; Schaefer, H. F. J. Chem. Phys. 2011, 135, 174107. (55) Ringer, A. L.; Sinnokrot, M. O.; Lively, R. P.; Sherrill, C. D. Chem.Eur. J. 2006, 12, 3821−3828. (56) Hohenstein, E. G.; Duan, J.; Sherrill, C. D. J. Am. Chem. Soc. 2011, 133, 13244−13247. (57) Stone, A. J. Chem. Phys. Lett. 1981, 83, 233−239.

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This material is based upon work supported by the National Science Foundation (Grant No. CHE-1011360) and the Center for Computational Molecular Science and Technology is funded through an NSF CRIF award (CHE-0946869).

■

REFERENCES

(1) Meyer, E. A.; Castellano, R. K.; Diederich, F. Angew. Chem., Int. Ed. Engl. 2003, 42, 1210−1250. (2) Č erný, J.; Kabelác,̌ M.; Hobza, P. J. Am. Chem. Soc. 2008, 130, 16055−16059. (3) Salonen, L. M.; Ellermann, M.; Diederich, F. Angew. Chem., Int. Ed. Engl. 2011, 50, 4808−4842. (4) Delgado, M. C. R.; Kim, E.; da Silva Filho, D. A.; Brédas, J. J. Am. Chem. Soc. 2010, 132, 3375−3387. (5) Sinnokrot, M. O.; Sherrill, C. D. J. Phys. Chem. A 2006, 110, 10656−10668. (6) Sherrill, C. D.; Takatani, T.; Hohenstein, E. G. J. Phys. Chem. A 2009, 113, 10146−10159. (7) Raju, R. K.; Bloom, J. W. G.; An, Y.; Wheeler, S. E. ChemPhysChem 2011, 12, 3116−3130. (8) Tsuzuki, S.; Honda, K.; Uchimaru, T.; Mikami, M.; Tanabe, K. J. Am. Chem. Soc. 2002, 124, 104−112. (9) Sinnokrot, M. O.; Valeev, E. F.; Sherrill, C. D. J. Am. Chem. Soc. 2002, 124, 10887−10893. (10) Janowski, T.; Pulay, P. Chem. Phys. Lett. 2007, 447, 27−32. (11) Pitoňaḱ , M.; Neogrády, P.; Ř ezác,̌ J.; Jurečka, P.; Urban, M.; Hobza, P. J. Chem. Theory Comput. 2008, 4, 1829−1834. (12) Wheeler, S. E.; Houk, K. N. J. Am. Chem. Soc. 2009, 131, 3126. (13) Janowski, T.; Ford, A. R.; Pulay, P. Mol. Phys. 2010, 108, 249− 257. (14) Janowski, T.; Pulay, P. Theor. Chem. Acc. 2011, 130, 419−427. (15) Hohenstein, E. G.; Sherrill, C. D. J. Chem. Phys. 2010, 132, 184111. (16) Casella, G.; Saielli, G. New J. Chem. 2011, 35, 1453−1459. (17) Denis, P. A. Chem. Phys. Lett. 2011, 516, 82−87. (18) Josa, D.; Rodriguez, O. J.; Cabaleiro, L. E. M. Phys. Chem. Chem. Phys. 2011, 13, 21139−21145. (19) Filatov, A. S.; Scott, L. T.; Petrukhina, M. A. Cryst. Growth Des. 2010, 10, 4607−4621. (20) Das, A.; Sood, A. K.; Maiti, P. K.; Das, M.; Varadarajan, R.; Rao, C. N. R. Chem. Phys. Lett. 2008, 453, 266−273. (21) Janowski, T.; Pulay, P.; Karunarathna, A. A. S.; Sygula, A.; Saebo, S. Chem. Phys. Lett. 2011, 512, 155−160. (22) Wu, Q.; Yang, W. J. Chem. Phys. 2002, 116, 515−524. (23) Grimme, S. J. Comput. Chem. 2004, 25, 1463−1473. (24) Grimme, S. J. Comput. Chem. 2006, 27, 1787−1799. (25) Grimme, S.; Antony, J.; Ehrlich, S.; Krieg, H. J. Chem. Phys. 2010, 132, 154104. (26) Becke, A. D.; Johnson, E. R. J. Chem. Phys. 2005, 122, 154104. (27) Becke, A. D.; Johnson, E. R. J. Chem. Phys. 2006, 124, 014104. (28) Becke, A. D.; Johnson, E. R. J. Chem. Phys. 2007, 127, 154108. (29) Johnson, E. R. J. Chem. Phys. 2011, 135, 234109. (30) Petrukhina, M. A.; Andreini, K. W.; Mack, J.; Scott, L. T. J. Org. Chem. 2005, 70, 5713−5716. (31) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. Rev. Lett. 1996, 77, 3865−3868. (32) Hehre, W. J.; Ditchfield, R.; Pople, J. A. J. Chem. Phys. 1972, 56, 2257−2261. (33) Becke, A. D. J. Chem. Phys. 1993, 98, 5648−5652. (34) Zhao, Y.; Schultz, N. E.; Truhlar, D. G. J. Chem. Theory Comput. 2006, 2, 364−382. (35) Zhao, Y.; Truhlar, D. G. Theor. Chem. Acc. 2008, 120, 215−241. (36) Kendall, R. A.; Dunning, T. H.; Harrison, R. J. J. Chem. Phys. 1992, 96, 6796−6806. 11926


Buckyplates and Buckybowls: Examining the Effects of Curvature on Ï

Recommend Documents

Buckyplates and Buckybowls: Examining the Effects of Curvature on Ï